[time-nuts] Have I wrecked my FE-5680A?

Bob Camp kb8tq at n1k.org
Fri Apr 29 23:20:04 UTC 2016


Hi

Just to toss one more possibility into the pile:

If there is an RTOS buried in the device, it’s probably got some sort of stack management processing 
going on in the background. Tasks get stacks and they are of a pre-determined size. (Yes, that’s a lot
of guessing already). If so, you *can* trigger some really wonky behavior by overrunning the max stack
limit. Most of the stuff I’ve seen does very little checking. It just happily goes off into wherever and never
comes back. If that “wherever” is the flash erase or eeprom write routine, watch out…

Bob


> On Apr 29, 2016, at 5:25 PM, Nick Sayer via time-nuts <time-nuts at febo.com> wrote:
> 
> 
>> On Apr 29, 2016, at 12:07 PM, Chuck Harris <cfharris at erols.com> wrote:
>> 
>> We have to sort of conclude that the failure is either
>> caused by bad data getting stored, or some sort of overflow
>> error.
>> 
>> It really is pretty unlikely that the firmware has been
>> changed, unless you happened to accidentally start a
>> flash burn routine, and it wasn't qualifying packets
>> properly.
>> 
>> If it isn't an overrun, or overflow error, then I would
>> suspect what you are sending to the port.  I have worked
>> with guys that are totally flustered when doing firmware
>> for devices that should qualify the data they receive.
>> When they don't know what to do (or are too lazy to try
>> and figure out the right thing to do) they simply stub
>> out the error path and go on.  Yuck!
>> 
>> If you have been sending an ill formed packet, you best
>> stop doing that ;-)
> 
> Oh, yeah, that’s fixed. While I’m waiting for a replacement 5680A my plan of record is to try and create a simulacrum out of an xCXO (the AVR of my board is clocked from the oscillator, so it won’t work without getting 10 MHz from something) and a serial port to make absolutely sure the data is correct before trying again.
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.




More information about the Time-nuts_lists.febo.com mailing list